Broad Ideas and Gory Details: The State of Health GIS
The first three quarters of the opening day of the Esri Heath GIS Conference held in Cambridge, MA on Oct 14th was the broad ideas part. It was very much a confirmation that GIS matters in the delivery of health care. Whether looking at nationwide data as the Dartmouth Atlas of Healthcare does, or worrying about the welfare of children as a variety of state and local professionals do or tackling a disparate population with a variety of needs as the State of Louisiana Department of Health and Hospitals does, or determining the best way for the Veterans Health Administration cares for its patients, GIS can help. When I asked a few attendees what they learned in these sessions they replied, in essence “Ummmmm, I’m looking forward to the breakout sessions with more details!” I understood the sentiment to mean that the sessions were interesting and informative, but that they’d not yet found something “to take home” yet.
What I found most interesting while watching the PowerPoints fly by in these broad-based sessions, is that the issues in GIS use in health areas are the same as just about every other industry. They include access to data, sharing data, convincing new users to look at a map, distinguishing between a 30 mile buffer and a 30 minute drive time polygon, leadership, technical capacity, training... The good news is that the community of GIS professionals, along with vendors and educators, are tackling those issues. This conference is in fact one way to address these challenges.
I learned one important set of idea. In fact, David Goodman, Professor of Pediatrics and of Health Policy at The Dartmouth Institute for Health Policy and Clinical Practice and a Co-Principle Investigator on the Dartmouth Atlas of Health Care, specifically said he wanted us to take away key ideas about unwarranted variation in health care use. I’m going to do my best to explain them below, but there’s a nice PDF I found that does a better job - and it’s by Goodman!
Unwarranted variation refers to variations in health care use across geographies that are not explained by people having a certain illness or preference to visit one medical center over another. No, these are related to health care system performance. What causes the variation? Per Godman, three kinds of care, evidence-based care, preference-based care, and supply sensitive care.
Evidence based care relates to health systems following “best practices” such as giving “beta blockers” right after (and apparently for some time after) a heart attack. When a best practice is new. only some systems follow them in a good percentage of cases. Over time, virtually all systems do. That’s currently the case with “beta blockers.”
Preference based care refers to systems that have a leaning to one action or another, but that patients may not be aware of or have a say in. For example, some hospitals are far more likely to have patients at the end of life spend their final days at home. Others are more likely to have patients spend them in the hospital. If a patient knew this, would they choose one over the other?
Supply sensitive care refers to Romer’s Law: A built bed is a full bed. Basically, if there is capacity for one kind of treatment or another in a geography it will be used. But it may not be where it’s really needed and, this increased capacity is not necessarily related to better outcomes!
The final breakout paper sessions of the day did indeed provide the gory details for the about 170 people in attendance. A standing room only crowd gathered to learn about an evolving study of where in terms of land use children get their physical activity. We learned about challenges of data capture (lame, but then state of the art GPSs - um, the one I currently use in my run training, a Forerunner 201), the challenges of middle school kids (who won’t stay outside in the cold until the GPS gets a fix), the challenges of land use classification (how do you define polygons of streets and sidewalks or automatically classify the GPS data into categories). I was please do t hear a bit shout out to MassGIS for providing great datasets for this study.
There was also a crowd, with pencils poised to write down website links, for a session on data collection about the U.S. health workforce from the National Center for the Analysis of Healthcare Data. Ann Peton, who was the state GIS coordinator for Iowa (and mentioned NSGIC several times during the morning session in her questions!) started this center herself and detailed how tough it was to get the data, what the center gives away and what it sells (to keep itself funded).
There were so few mentions of Esri or its products during the majority of the day that an attendee might even think he or she was not at an Esri conference! But of course it was an Esri event, so to break up the “broad ideas” and “gory details” sessions of the day, Bernie Szukalski provided the Esri vision with a focus on the ArcGIS Online platform. He demoed how easy it is to build an app but also shared a vision for the current state of GIS use. My interpretation of his graphic is below. It details, more broadly that in healthcare I believe, who needed GIS vs. who actually had access to and use of it.